Zoho chief executive officer Mani Vembu said the company is doubling down on vertical SaaS, identifying it as a "major growth area similar to where horizontal SaaS was two decades ago".
Speaking to Moneycontrol in an interaction, Vembu said Zoho plans to take its learnings from horizontal SaaS to specific industries like auto dealership management (Auto DMS) and BFSI, aiming to drastically cut implementation timelines from 18 months to just six. Zoho is now productising its early wins in these verticals for scale and expansion.
“Many vertical SaaS players don’t even publish their pricing, which shows the market is ripe for disruption,” he said.
The company’s newly launched Zia LLM, a homegrown enterprise-grade large language model trained on 7 billion parameters, is at the core of this approach. It’s not designed to answer trivia or write poems, it’s trained on business documents and tailored to Zoho’s applications across functions such as sales, marketing, support, and finance.
Alongside Zia LLM, the company has also rolled out Zia Agents, task-specific AI agents built into Zoho’s suite of applications. These agents automate routine tasks such as drafting emails, creating reports, summarising support tickets, and scheduling meetings, while allowing human oversight.
Also Read: Zoho launches Zia LLM, a homegrown enterprise AI model
In a recent report by AI startup accelerator Upekkha, global enterprise spending on vertical AI, industry-specific artificial intelligence systems, is projected to surge from $5 billion in 2024 to $47 billion by 2030, growing at a compound annual growth rate (CAGR) of over 40 percent.
Zoho is one of the few Indian SaaS companies that went global and horizontal early, offering dozens of tightly integrated software tools for enterprises, all built in-house. Now, it’s applying the same philosophy to vertical AI.
Ramprakash Ramamoorthy, Director of AI Research at Zoho, said the company’s models are designed with "purposeful intelligence" rather than general-purpose brilliance. “We are investing in modular, task-specific models. These are not just chatbots. They’re context-aware agents that understand enterprise data, workflows, and business logic,” he said.
While many AI companies rely on APIs from OpenAI or Anthropic to power their products, Zoho has taken the long road, training its own models, running them on its own data centers, and integrating them across its full-stack suite. This not only gives Zoho tighter control over latency and reliability, but also strengthens data privacy, a key concern for enterprise customers.
“In our stack, we control everything from the data layer to the model to the user interface. That’s what allows us to go deep,” said Ramamoorthy.
Vembu said the company is not looking to release standalone AI products or monetise the LLM itself. Instead, the model will serve as a “common engine” across Zoho’s suite of over 50 applications, with each app having agents that serve specific use cases.
Zoho has also started experimenting with retrieval-augmented generation (RAG) to combine the power of its models with structured enterprise data. In internal testing, the company has found that combining RAG with its task-specific agents improves accuracy and usability for business users.
“We don't need to compete with ChatGPT,” said Vembu. “That’s not our goal. Our goal is to solve for specific business needs—and we believe we are in a strong position to do that.”
Discover the latest Business News, Sensex, and Nifty updates. Obtain Personal Finance insights, tax queries, and expert opinions on Moneycontrol or download the Moneycontrol App to stay updated!
